Multi-material additive manufacturing: gradient structures and dissimilar alloy joining

Most metal AM builds one alloy per build. Multi-material AM — depositing two or more alloys in a single structure — is one of the most promising capabilities the technology can offer, and one of the least mature in commercial practice. This article explains what is currently achievable, where the fundamental materials science constraints lie, and what the realistic engineering applications are.

Why multi-material AM matters

The engineering motivation is straightforward: many high-performance components need different properties in different regions. A turbine disc needs oxidation resistance at the rim and fracture toughness at the bore. A biomedical implant needs bone-ingrowth porosity at the surface and fatigue resistance in the stem. A heat exchanger needs thermal conductivity in the channel walls and strength in the flanges.

Conventional manufacturing addresses this by joining dissimilar materials — brazing, welding, diffusion bonding, or mechanical fastening. Each approach introduces an interface that is typically the weakest point in the assembly, requires additional operations, and limits the geometry of the transition zone.

Multi-material AM offers a fundamentally different approach: the transition between alloys can be gradual (functionally graded material, FGM) rather than abrupt. A well-designed graded interface distributes the thermal and mechanical mismatch stress over a finite volume, eliminating the stress concentration at a sharp bimetallic joint.

Process options that enable multi-material AM

Not all AM processes can deposit multiple materials. The options differ substantially in capability:

Laser DED with dual powder hoppers

This is the most capable and most studied multi-material approach. LP-DED systems with 2–4 independent powder hoppers can vary the feed ratio between hoppers in real time, creating a composition gradient layer-by-layer.

Ti-6Al-4V to 316L stainless steel is the most-studied transition. The direct Ti/Fe interface is problematic (discussed below), but a graded path via vanadium interlayer — Ti-6Al-4V → Ti-V alloy → pure V → V-Fe alloy → 316L — has been demonstrated at Penn State (Beese et al., 2019) and achieves tensile strengths approaching the weaker alloy. The full transition zone is typically 5–20 mm thick.

IN718 to copper alloy (CuCrZr) is an aerospace heat exchanger application: the structural section is nickel superalloy, the heat transfer section is copper alloy. Direct joining by conventional methods is extremely difficult due to the melting point difference (1260°C for IN718 vs. 1080°C for Cu). LP-DED graded deposition bridges this gap, depositing IN625 → IN625/Cu blend → CuCrZr in controlled steps.

Parameters for dual-hopper DED:

• Hopper switch time: 5–30 s (mechanically limited by the powder delivery system)
• Minimum transition length per composition step: 1–3 mm (one to a few layers)
• Powder mixing ratio accuracy: typically ±2–5% by weight

Wire-arc (WAAM) with wire switching

WAAM systems with two wire feeders can switch between wires mid-build. Wire switching is slower than powder ratio adjustment — typically 30–60 s for a clean transition — and the transition zone is 3–10 mm rather than sub-millimetre. But WAAM wire switching is mechanically simpler and more reliable than dual-powder systems, and wire alloys are more widely available.

Demonstrated WAAM multi-material combinations include:

• Low-alloy steel (ER70S-6) to 316L stainless (ER316L)
• Aluminium 4043 to 5356 (limited mixing zone)
• Titanium CP Grade 2 to Ti-6Al-4V

The limited resolution of WAAM means it is suited to large, coarse transitions — structural zones rather than fine functional gradients.

Binder jetting with multi-material powder beds

Current commercial binder jetting systems (Desktop Metal, ExOne, HP Metal Jet) use a single powder bed and cannot vary composition spatially within the green part — binder jetting is effectively single-material for metals. Multi-material binder jetting is an active research area, but commercial capability is limited to multi-material sand casting moulds and some ceramic/polymer systems where coarse spatial resolution is acceptable.

FDM with dual extrusion

Dual-extrusion FDM (Stratasys, Bambu Lab, Prusa) can print two polymer materials in one build. This enables:

• Rigid/flexible combinations (ABS + TPU)
• Soluble support materials (PVA or HIPS alongside the structural polymer)
• Two-colour/two-material functional parts

FDM multi-material is mature and widely available, but is polymers only — the thermal and bond strength of polymer-polymer interfaces is acceptable for many applications, and the processing is straightforward.

LPBF: currently not multi-material

LPBF cannot deposit multiple alloys in a single build with current commercial systems. The process uses a uniform powder bed; once spread, the layer composition is fixed. Research systems with multiple powder dispensers exist (at Fraunhofer ILT, TU Wien, and others), but no commercial system provides this capability as of 2026. If you need multi-material metal AM today, the answer is DED.

The alloy compatibility problem

Not all alloy combinations can be joined by graded deposition. The fundamental constraint is intermetallic formation at the interface.

When two metals mix in the liquid state and solidify, the equilibrium phase depends on composition and temperature. If the equilibrium phase diagram contains brittle intermetallic compounds (IMCs) in the composition range of the mixed zone, the interface will be brittle regardless of how carefully the gradient is designed.

The titanium–steel interface is the canonical example. The Ti-Fe binary system contains the intermetallics TiFe and TiFe₂. Both are hard, brittle, and form within a narrow composition range around 50 at% Fe. Any direct Ti/steel mix passes through this composition range during solidification. As-deposited Ti-6Al-4V/316L structures with a direct interface fracture in the mixed zone at stresses far below either alloy's yield strength — the interface is essentially ceramic in character.

The vanadium interlayer solution (Beese group, Penn State, 2019) works because vanadium is fully miscible with both titanium and iron, and the Ti-V and V-Fe systems do not contain brittle IMCs over the relevant composition range. The graded path Ti-6Al-4V → Ti-V → V → V-Fe → 316L avoids the brittle composition window entirely.

The Kirkendall effect is a secondary concern at multi-material interfaces. During high-temperature post-processing (stress relief, HIP, annealing), atomic diffusion across a gradient interface is asymmetric — faster-diffusing species create vacancy supersaturation on their departure side, leading to Kirkendall porosity. For Ti/V/steel, Kirkendall porosity has been observed after HIP at ≥920°C. This constrains post-processing choices for multi-material structures.

Key incompatible pairs:

Design rules for gradient structures

A graded interface is not simply "mix the two alloys 50:50 in the middle." Proper FGM design requires:

Transition zone width: The minimum transition zone width to avoid stress concentration is governed by the thermal expansion mismatch and the elastic modulus gradient. For Ti-6Al-4V to 316L (CTE difference ~5 µm/m·K), a transition zone of ≥5 mm is generally required to keep interface stresses below yield in the cooler alloy. Thinner transitions concentrate stress and reduce fatigue life.

Step gradient vs. continuous gradient: Step gradients (discrete composition jumps every 1–2 mm) are easier to implement with powder hoppers and easier to characterise with EBSD and EDS. Continuous gradients are more effective at distributing mismatch stress but require smooth hopper ratio control. For most applications, a 5–10 step gradient over 5–15 mm is a practical compromise.

Thermal stress analysis: Always model the residual stress state from processing before committing to a gradient design. A zone that transitions from a low-CTE alloy (Ti: 8.6 µm/m·K) to a high-CTE alloy (316L: 16 µm/m·K) will have significant residual stress after cooling from deposition temperature. FEA with temperature-dependent properties is mandatory for safety-critical applications.

Geometry at the interface: Curved interfaces distribute stress more uniformly than flat ones. Where possible, orient the gradient interface perpendicular to the primary load direction — this minimises mode-I (opening) stress at the interface and shifts the failure mode toward shear, which gradient structures handle better.

Current engineering applications

Aerospace structural brackets (Ti → SS): The strongest commercial motivation is replacing titanium-to-steel mechanical fasteners with integral gradient joints. A bracket that is titanium at the airframe attachment and steel at the engine fitting, with a V-interlayer gradient in between, eliminates the bolted joint, reduces part count, and removes the fretting fatigue risk. Airbus and Boeing research programs have demonstrated this concept; no production application had been approved as of early 2026.

Turbine blade platforms (IN738 → IN625): Turbine blade platforms experience different temperature and stress profiles from the aerofoil. Graded composition between the platform and aerofoil — from a higher-toughness alloy to a higher-oxidation-resistance alloy — is achievable by LP-DED. Siemens Energy and GE Aerospace have active R&D in this area.

Biomedical implants (Ti-6Al-4V → CoCr): Cementless implants need bone-in-growth porosity (60–80% porosity, 200–500 µm pore size) at the bone-contacting surface and solid metal for load-bearing sections. Multi-material LP-DED can deposit a graded lattice-dense transition, although currently EBM and LPBF with designed porosity are the more common manufacturing routes. The Ti → CoCr gradient is metallurgically feasible — both alloys are biocompatible and the Ti-Co system has limited IMC formation.

Heat exchangers (CuCrZr → IN718): For rocket engine applications requiring high-pressure, high-temperature operation with high heat flux, copper-alloy cooling channels within a structural nickel alloy body is the target architecture. Several agencies (DLR, NASA) have demonstrated DED fabrication of this combination.

Characterisation challenges

Multi-material AM parts require characterisation at multiple scales:

Interface microstructure: Electron backscatter diffraction (EBSD) maps across the gradient zone reveal grain structure evolution, phase identification, and the presence of IMCs. Wavelength-dispersive X-ray spectroscopy (WDS) quantifies local composition at submicron resolution. This is slow and expensive — a full gradient characterisation by EBSD/WDS can take several days per specimen.

CT scanning for delamination: The highest-risk failure mode is delamination at the interface — either immediate (from thermal stress during processing) or fatigue-driven. Industrial CT scanning at 5–50 µm voxel resolution is the standard non-destructive check. The interface appears as a density gradient in CT; a clear step or discontinuity indicates a problem. CT is particularly useful for detecting Kirkendall porosity that develops after HIP.

Mechanical testing across the interface: Standard tensile specimens cannot capture the interface performance if the specimen gauge section does not include the interface. Dog-bone specimens must be oriented to place the interface at mid-gauge, which usually means machining from specific regions of the deposit. Fracture toughness testing (compact tension, SE(B)) is required for safety-critical qualification; the crack must be oriented to propagate through the interface zone.

Maturity and limitations

Multi-material DED is best described as TRL 4–6 (laboratory to limited demonstration): demonstrated in subscale hardware, characterised in academic literature, but not yet in serial aerospace production.

The key limitations that prevent wider adoption:

Limited commercial availability. Fewer than a dozen LP-DED systems worldwide have multi-hopper configurations with the controls needed for repeatable gradient deposition. Most are at research institutions or OEM internal labs.

Qualification framework gaps. ASTM F3187 and AWS D20.1 cover single-material DED. There is no established qualification standard for graded structures. Aerospace primes developing multi-material DED parts are writing their own qualification plans, which is expensive and slow.

Process repeatability. Powder flow rate from individual hoppers varies by ±2–5% between runs due to powder packing changes and nozzle wear. Over a 20 mm gradient zone, this is typically acceptable; over a 2 mm sharp transition, it is not.

Post-processing constraints. As noted above, HIP temperatures that are standard for single-alloy DED may cause Kirkendall porosity in certain gradient combinations. Heat treatment optimisation for multi-material structures is an open problem and must be solved case-by-case.

For most engineering applications today, if you need a multi-material structure, the realistic options are: (1) LP-DED with a characterised gradient for Ti/V/steel or Ni/Cu combinations, (2) conventional joining (diffusion bonding, brazing) for other combinations, or (3) accept the interface by design and engineer the structure to keep interface stresses below threshold.

Further reading

• Material substitution tool — compare property profiles when evaluating alloy transitions
• DfAM checklist — design rules including multi-material considerations
• Cost per part — model the economics of DED vs. conventional joining approaches